Start using components with $ref

[A5rocks]

This change as far as I can tell is backwards compatible due to the overloads, though it does add a deprecated case.

I have not tested this at all so it probably doesn't work. Just making this to signal intent and solicit early feedback. Fixes #108.

EDIT: tests are now made. I think that showcases what this PR does. Pretty sure it's not exhaustive but not sure what other tests you might want!

[nx-cloud]

☁️ Nx Cloud Report

CI is running/has finished running commands for commit 75c66b2. As they complete they will appear below. Click to see the status, the terminal output, and the build insights.

📂 See all runs for this branch

---

✅ Successfully ran 2 targets

• nx affected:test --base=origin/main --codeCoverage
• nx affected:lint --base=origin/main

---

Sent with 💌 from NxCloud.

[Brian-McBride]

I pulled in another PR that updates the underlying open API schema lib. I think you'll need to update this PR as well.

[A5rocks]

@Brian-McBride hi! So, everything's typed as OpenAPI 3.1 schemas meaning e.g. no "nullable" field. I actually have a change locally that in theory updates everything and was about to push it when I realized that's probably a bit too much of a change (and I've got one test I can't really figure out, plus it somehow breaks a feature...).

Basically, should I:

• push a ton of changes to use openapi 3.1 as the types indicate
• switch the types to openapi 3.0 (and then keep those changes to make a proper PR that can be in a minor or major release).

[A5rocks]

I'll just change back types to openapi 3.0 for now...

[Brian-McBride]

I am sorry, I really need a co-maintainer.

It looks fine. RIght now, everything is at 3.1 right?
The concern is the tests that go from exclusiveMinimum: 0 to exclusiveMinimum: true

I'll pay attention to the repo close here for the next few days. I'd suggest we bring it to he "current" standards. I just did a major release because the peer dep is a major release. We could do a another major?

But again, a test that goes from 0 to true seems off. I could see 0 going to false and not breaking downstream code.

[A5rocks]

It looks fine. RIght now, everything is at 3.1 right?

Err, no. At the moment, main branch is a strange combination of 3.0 and 3.1. For instance, it still uses nullable. I had a decently large diff to make everything match 3.1 behavior but I decided it would probably be best out of this PR.

I think I have that stashed somewhere, if you want it as another (or in this?) PR.

[DHFW]

Hello @A5rocks @Brian-McBride, what is the status of this? I would really appreciate if this goes to main. Thank you!

[A5rocks]

Sorry about the wait; I'll see if I can find my stash that moves things to 3.1 and make that a separate PR (after fixing it up and bringing it up to date)

[DHFW]

@A5rocks sorry man, did you had the time to work on this? Looking forward to it. Thanks!

[A5rocks]

Sorry, I put this off then forgot. I've now made #173 for that OpenAPI 3.1 support.

Thank you for reminding me!